390 



Professor Sir James Dewar 



[Jan. 18, 



out for five minutes, the loss of Imovancj produced the sag shown 

 in the second drawing («), taken when rupture was imminent between 

 the first and second segments : the measurements could not, there- 

 fore, be completed. The ratio of length to breadth of the longest 

 segment (on the point of instability) is 1*17 to 1. Comparative 

 measures are also shown in the two drawings {b) of the considerable 

 contraction that takes place in an air-blown column V»etween its 

 •early condition when uniformly coloured, with all superfluous hquid 



(«) 



(^) 



cm" 



-80- 



n- 



-^ 



5Qr- 



—m- 



—3b- 



m-- 



po\tvt 



7.H 



Fig. 28. 



removed, and its final condition of complete blackness, and therefore 

 minimum mass. The successive heights of the segments in a column 

 half an hour after blowing (uniformly coloured, no excess liquid) 

 were, counting from the top, 5*1 to 5*2 cm. for the first four ; then 



m succession, 5*0, 4-8, 4-3, 4*2, 4 



2 — at which point 



contact was made with a larger segment on a fixed support. 



The long glass cylinder in Fig. 6 was chiefly used for these 

 measurements. The sliding support rod was ada])ted for manipu- 



